A COMPARITIVE STUDY ON FAKE JOB POST PREDICTION USING DIFFERENT DATAMINING TECHINIQUES

Abstract

The proliferation of online job boards and
other forms of mass communication has
made posting openings for new positions a
routine occurrence in today's society.
Therefore, everyone should be quite worried
about the problem of predicting bogus job
postings. False job posing prediction has
many of the same difficulties as other
categorization problems. A variety of data
mining and classification algorithms,
including KNN, decision trees, support
vector machines, naive bayes, random forest,
multilayer perceptrons, and deep neural
networks, are suggested in this study as ways
to determine the authenticity of a job posting.
We conducted experiments using the 18000-
sample Employment Scam Aegean Dataset
(EMSCAD). This classification challenge is
well-suited to deep neural networks as
classifiers. This classifier for deep neural
networks is built with three thick layers.
When it comes to predicting a fake job
posting, the trained classifier demonstrates
about 98% DNN classification accuracy.